Inter-cell interference (ICI) is a serious problem in multi-cell orthogonal frequency division multiple access systems and should be mitigated by a smart radio resource management scheme. The authors present here a novel downlink adaptive resource allocation (RA) algorithm based on ICI measurement to achieve both multi-user diversity gain and the cancellation of ICI. The proposed algorithm is performed in two steps: first, interference management by user grouping into different clusters and second, subchannel allocation in order to maximise the total throughput of the network. It does not need any priori frequency planning and no precise signal-to-interference plus noise ratio information is required for subchannel allocation. In both steps of the algorithm, the authors propose a graphic framework and use Hungarian method to solve subproblems. Simulation results indicate that the authors algorithm mitigates the ICI successfully and can improve the total throughput of network by 10-15% compared with existing adaptive RA schemes.